Information Technology Reference
In-Depth Information
Disk 0
Disk 1
Data Block 0
Data Block 1
Data Block 2
Data Block 3
Data Block 0
Data Block 1
Data Block 2
Data Block 3
Data Block 4
Data Block 5
Data Block 6
Data Block 7
Data Block 8
Data Block 9
Data Block 10
Data Block 11
Data Block 12
Data Block 13
Data Block 14
Data Block 15
Data Block 16
Data Block 17
Data Block 18
Data Block 19
Data Block 4
Data Block 5
Data Block 6
Data Block 7
Data Block 8
Data Block 9
Data Block 10
Data Block 11
Data Block 12
Data Block 13
Data Block 14
Data Block 15
Data Block 16
Data Block 17
Data Block 18
Data Block 19
Figure14.5: RAID 1 with mirroring.
14.2.2
RAID: Multi-disk redundancy for error correction
Given the limits of physical storage devices, storage systems use additional
techniques to get acceptable end-to-end reliability. In particular, rather than
trying to engineer perfectly reliable (and extremely expensive) storage devices,
storage systems use Redundant Arrays of Inexpensive Disks (RAIDs) so that a
partial or total failure of one device will not cause data to be lost.
Basic RAIDs
A Redundant Array of Inexpensive Disks (RAID) is a system that spreads data
Definition: RAID
redundantly across multiple disks in order to tolerate individual disk failures.
Note that the term RAID traditionally refers to redundant disks, and for sim-
plicity we will discuss RAID in the context of disks. The principles, however,
apply equally well to other storage devices like flash drives.
Figures 14.5 and 14.6 illustrate two common RAID architectures: mirroring
and rotating parity.
Mirroring. In RAIDs with mirroring (also called RAID 1 ), the system
Definition: mirroring
Definition: RAID 1
writes each block of data to two disks and can read any block of data from
either disk as Figure 14.5 illustrates. If one of the disks suffers a sector or
whole-disk failure, no data is lost because the data can still be read from
the other disk.
